Fast and Accurate Robot Localization through Multi-Layer Pose Correction

Localization is a key issue for autonomous navigation of mobile robots. Generally, the location algorithm based on a single sensor has poor adaptation to complicated environments, which results in large localization error of robots. In this paper, we study the data fusion of odometry, Inertial Measure Unit (IMU) and two-dimensional (2D) laser radar. The multi-sensor data is gradually integrated together with the correction of robot pose, by means of Unscented Kalman Filter (UKF), Augmented Monte Carlo Localization (AMCL) and 2D Distribution-to-Distribution (D2D) Normal Distributions Transform (NDT), for quickly improving the localization accuracy of robot. Experimental results show that this method enables the robot to locate its position quickly and accurately.